Drug-drug interactions (DDIs) are a significant clinical and public health burden, especially for older adults. This burden will increase further as the population ages and the degree of polypharmacy increases. Knowledge about DDIs comes from many sources, with rigorous population-based studies of clinical outcomes providing the most clinically useful information. Because very few population DDI studies have been performed, there is profound disagreement among the DDI compendia about which drug pairs cause clinically important DDIs. Further, uncertainty about which potential DDIs are clinically important has hampered efforts to mitigate the effects of DDIs, such as computerized decision support. Population studies of the clinical effects of DDIs are therefore badly needed, and should be integrated with research on the biological basis of DDIs. In this renewal application, we propose to extend our approach to population DDI studies in two important and innovative ways: 1) by actively and explicitly incorporating mechanistic considerations; and 2) by implementing a sequential approach of initially identifying and quantifying risk, then, where warranted based on this step, characterizing important DDIs further, with the goal of providing the basis for future efforts to mitigate risk. Based on an extensive and structured consultative process to identify research foci, we propose to conduct a series of population DDI studies using a large 5-state Medicaid/Medicare dataset. These studies will examine hypothesized interactions involving lipid lowering drugs, anti-clotting drugs, and oral anti-diabetic drugs, three drug classes with enormous relevance for older adults, and with high potential for causing clinically important DDIs. Thus, we will study potential DDIs of great clinical and public health importance, especially to older adults, and in the process develop data that will help elucidate the mechanisms of these potential DDIs.